Data Science - Machine Learning

• Data Preprocessing
• Data Science Life Cycle (DSLC)
• Jupyter Notebook

Data Preprocessing

• Basic
  • For higher accuracy
• Normalization (Min-Max Normalization)
  • Scaling technique that rescales data of features in the range [0, 1]
  • X^' = X-min/max-min
  • Code

        from sklearn import preprocessing
        scaler = preprocessing.MinMaxScaler(feature_range=(0, 1))
        rescaled = scaler.fit_transform(data)
    

• Standardization (Z Score Normalization)
  • Scaling technique that rescales data such that they will have properties of standard normal distribution with mean (μ=0), standard deviation (σ=1)
  • Z = X-μ/σ
  • Code

        from sklearn import preprocessing
        scaler = preprocessing.StandardScaler().fit(data)
        rescaled = scaler.transform(data)
    

• Binarization
  • Scaling technique that rescales data into "0" or "1"
  • Used when dataset contains probabilities
  • Code

        from sklearn import preprocessing
        scaler = preprocessing.Binarizer(threshold=0.5)
        rescaled = scaler.transform(data)
    

Data Science Life Cycle (DSLC)

• Business Requirement (BR)
  • Deciding the objectives/goals
• Data Acquisition (DA)
  • Collecting data from Source that suits your Objective
• Data Processing (DP)
  • Making raw data suitable for model training
• Data Exploration (DE)
  • Understand data
• Modelling
  • Choosing a model and applying
• Deployment
  • Deploy in an environment
  • Technology
    • Backend => Django and Flask
    • Docker and Kubernetes
    • Streamlit

Jupyter Notebook

• Installation
  • pip install jupyter => Install Jupyter
  • pip install notebook => Alternate way
• Launch Notebook
  • jupyter notebook => Opens a server in browser, Runs notebook
  • Click New > Python 5 => Open a new Notebook
• Shortcuts
  • enter => New line
  • shift + enter => Run
  • tab => Suggestion/Autocomplete
• Commands
  • !mkdir name => Create file
  • %lsmagic => Shows magic commands
  • %%HTML => To write html
  • %matplotlib inline => To write matplotlib commands